2. cancer in humans...2. cancer in humans 2.1 introduction 2.1.1 general aspects diesel and gasoline...

2. CANCER IN HUMANS

2.1 Introduction

2.1.1 General aspects

Diesel and gasoline engine exhausts have been evaluated previously in the IARC Monographs (IARC, 1989). Since that time, a large number of cohort and case–control studies have been published on the topic, many of which provided high-quality data on exposure and potential confounding factors. General aspects on the methodological advantages and limitations of the various study designs are discussed briefly below, including a discussion of developments in the methods of exposure assessment and problems regarding bias and confounding.

Occupational cohort studies typically provide reasonably accurate data on current exposure levels, but frequently lack information on historical exposure levels and other risk factors, such as tobacco smoking habits. Over the past few decades, the quality of cohort studies has been further improved by the inclusion of more detailed data on exposure and more advanced exposure modelling, and some now incorporate individual data on tobacco smoking. However, full lifetime occupational histories are often unavailable, which reduces the ability to adjust the findings for occupational exposures incurred during employment outside the industry under study. Comparisons are made between groups with different levels of exposure, either within or outside the study population. In general, the use of an internal unexposed group gives more

valid risk estimates than that of an external group, because less bias is introduced from the incomparability of lifestyle-associated factors, such as smoking, and the general health status of the exposed and unexposed groups. Studies that investigate exposure–response associations with measures of exposure to diesel exhaust (e.g. duration, average exposure and/or cumulative exposure) have been given greater weight in the evaluation of carcinogenicity.

A specific group of cohort studies are based on record linkage, and link routinely collected population data on occupational titles to national registers of cancer incidence or mortality. These studies usually provide very crude data on exposure that typically consist of a job title in a specific year, no data on tobacco smoking and no lifetime occupational histories, and are consequently usually viewed as generating hypotheses. This type of study has been considered to give only supportive evidence for the present evaluation.

Finally, some cohort studies use proportionate mortality ratios, a methodology that is applied when data are available on deaths for those employed at a specific industry, but not for the total population at risk, and the distribution of causes of death in the study population is compared with that of an external comparison group. Proportionate mortality ratio studies may give biased results, because it is not possible to assess whether an observed excess in the proportion of deaths from a specific cause is due to a true excess of risk or a reduced risk of deaths for

147

IARC MONOGRAPH – 105

other causes. For this reason, and in view of the large number of high-quality studies available, proportionate mortality ratio studies have not been considered for the present evaluation.

Case–control studies are sometimes considered to be sensitive to inaccurate recall of previous exposures, with the potential for biased findings, which is a particular concern when information on exposures is based on self-assessment. This potential bias is largely reduced when exposure assessment is based on the assignment of exposures from a lifetime occupational history by an expert assessment method or the application of a job–exposure matrix (JEM), or combinations of the two. The strength of case–control studies is that researchers can obtain detailed individual data on tobacco smoking and other potential risk factors, as well as lifetime occupational histories, from the study subjects. These studies may be based on the general population or be nested within defined occupational cohorts. A potential problem in the former is that detailed exposure data are rarely available, resulting in potential non-differential exposure misclassification, which tends to attenuate the observed risks. A design that incorporates the advantages of detailed data on exposure and individual data on risk factors is the nesting of a case–control study within a cohort. This type of study may provide very high-quality information on the association of occupational exposures with cancer. Finally, case–control studies may recruit controls directly from the general population or from individuals with diagnoses other than that under study. Controls recruited randomly from the population are considered to reflect the prevalence of exposure in the study population more accurately than hospital controls, and afford less potential for bias.

For the present evaluation, the Working Group identified three broad types of study, in order of increasing quality: (1) studies that reported risk for cancer by occupation, with no reference to whether or not the occupation

indicated exposure to motor exhaust. Because the predictive value of an occupational title may differ considerably between study settings and countries, such studies have not generally been included for the evaluation of lung cancer. However, for other cancer sites, for which fewer data were available, such studies have been included; (2) studies that reported risk for cancer on the basis of occupational titles, in which the titles were used to indicate potential or definite exposure to diesel or gasoline exhaust. Frequently, duration of employment in such jobs was used as a proxy for quantitative exposure. All such pertinent studies have been reviewed; (3) studies that aimed to assess individual exposure to motor exhaust (diesel, gasoline or both) quantitatively through measurements, model-ling, expert assessments, JEMs or other means. In many cases, these studies investigated the intensity and duration of exposure, in addition to cumulative exposure, and had access to full occupational and smoking histories. These studies have been given the greatest weight in the present evaluation.

2.1.2 Aspects of exposure assessment methods

Diesel exhaust is a complex mixture of variable characteristics and several markers of exposure have been used, including polycyclic aromatic hydrocarbons (PAHs), nitrogen dioxide, nitrogen oxides, particulate matter (PM) and elemental carbon (EC). Any proxy of diesel exhaust may or may not reflect accurately one of its underlying carcinogenic components. For instance, in the studies reviewed that investigated proximity to diesel emission sources, EC was a good marker for diesel exhaust. However, this does not imply that EC is, or that it accurately reflects, the causal agent. Therefore, the association between such markers and the true substances of etiological interest may differ to some extent by time and place.

148

Diesel and gasoline engine exhausts

The present evaluation covers both diesel and gasoline exhaust. Many occupations that entail exposure to engine exhausts, for example professional drivers and some garage workers, involve exposure to a mixture of the two exhausts. Cohort studies that more specifically addressed exposure to diesel exhaust were based on railroad workers, miners and bus garage employees. Cohorts of professional drivers often comprised individuals with combined exposure to diesel and gasoline exhausts. Several population-based case–control studies that used expert assessment or JEMs to classify exposure presented separate risk estimates for each of the two exhausts. However, in population-based studies, few occupations generally entailed exposures to specific types of engine exhaust and few individuals incurred very high exposures. Some collinearity may have also arisen between exposures to diesel and gasoline engine exhausts. The main body of available evidence was related to diesel exhaust, while the data that specifically addressed the risk of cancer from exposure to gasoline exhaust among individuals with no concurrent exposure to diesel exhaust were very limited.

2.1.3 Studies of environmental air pollution

Several studies showed associations between lung cancer and ambient air pollution. In addition, exposure to specific components of air pollution, for example PM2.5, has been linked to lung cancer (Samet & Cohen, 2006; EPA, 2009). Ambient air pollution comprises emissions from vehicles fuelled by diesel and gasoline, but also those from a variety of other sources and processes, including industrial air pollution. At present, it is very difficult to assess the specific contributions of these sources to the observed cancer risk. These studies have not been included in the review, because they would contribute little information in addition to the studies reviewed here.

2.2 Cohort studies

See Table 2.1

2.2.1 Railroad workers

Howe et al. (1983) conducted a cohort study of 43 826 male pensioners of the Canadian National Railway Company who had retired before 1965 and who were alive at the start of that year, as well as those who retired between 1965 and 1977. The cause of death of 17 838 pensioners who died between 1965 and 1977 was ascertained by computerized record linkage to the Canadian national mortality database. A total of 76 different occupations were represented in the cohort. Experts classified each occupation in terms of exposure to diesel fumes (unexposed, possibly exposed and probably exposed), coal dust (unexposed, possibly exposed and probably exposed) and any other fumes or substances. The analyses compared the mortality of different groups of railroad workers with that of the Canadian population and then compared those who presumably incurred higher exposures with those who presumably incurred lower exposures, by calculating their standardized mortality ratios (SMRs). A total of 933 lung cancer deaths resulted in an overall standardized mortality ratio of 1.06. A comparison of the possibly and probably exposed with the unexposed (239 lung cancer deaths) provided a relative risk (RR) for lung cancer of 1.20 (P = 0.013; 407 lung cancer deaths) and 1.35 (P

Table 2.1 Selected cohort studies of cancer in railroad workers, bus garage workers, professional drivers, miners, heavy equipment operators and other workers exposed to diesel exhaust

IARC M

ON

OG

RAPH

– 105

Reference Total No. of Exposure assessment Organ site Exposure No. of Relative risk Covariates Location, subjects (ICD code) categories exposed (95% CI) Comments follow-up cases period

Railroad workers Howe et al. 43 826 retired Review of job titles by Trachea, SMR Men; Canadian mortality records; (1983) Canada,

railroad workers experts employed by railroad

bronchus and lung

All railroad workers

933 1.06 (P > 0.05) no adjustment for tobacco smoking

1965–77 Possibly exposed 407 1.20 (P = 0.013) Probably exposed 279 1.35 (P

Table 2.1 (continued)

Reference Location, follow-up period

Total No. of subjects

Exposure assessment Organ site (ICD code)

Exposure categories

No. of exposed cases

Relative risk (95% CI)

Covariates Comments

Garshick et al. Train crew work (2004) US Railroad Workers cohort, 1959–96 (cont.)

Garshick et al. (2006) US Railroad Workers cohort, 1959–96

39 388 Industrial hygiene review of job history and exposure sampling

Lung (162)

(yr; 5-yr lag) 0–



Laden et al. 52 812 Industrial hygiene Lung Train crew work

(2006) review of job history (yr; 5-yr lag) US Railroad and exposure Hired 1939–44 Workers sampling Unexposed cohort,

1959–96 0–



Guo et al. (2004a) Finland, records from 1971–95

Finnish working population, census data in 1970; 67 121 men

Longest held job in 1970 census

Lung Locomotive driver

85 SIR 0.63 (0.51–0.78)

All working Finns born after 1906; linkage to Finnish Cancer Registry; indirect adjustment for tobacco smoking

Guo et al. (2004b)

Urinary bladder

Locomotive driver

22 SIR 0.85 (0.53–1.28)

Same census linkage study as Guo et al. (2004a); cases in men

Finland, 1971–95 Bus garage workers Gustavsson et al. (1990)

695 bus garage workers, 1945–70

JEM for diesel exhaust exposure;

Lung Entire cohort 17 SMR 1.22 (0.71–1.96)

Expected numbers based on Stockholm rates; occupationally

Stockholm, intensity × duration active men; no information on Sweden, score tobacco smoking; exposure to mortality, asbestos assessed; no effect of 1952–86; exposure on lung cancer cancer incidence, 1958–84

Nested case– control study: 20 incident cases/120

Diesel exhaust score

OR

Controls matched on age (± 2 yr)

controls 0–10 5 1.00 (referent) 10–20 2 1.27 (0.21–7.72) 20–30 3 1.56 (0.34–7.16) > 30 10 2.63 (0.74–9.42)

Professional drivers Bus drivers Balarajan & McDowall

3392 professional drivers

Job title in 1939 census and still alive in 1950

Lung Bus and coach drivers

18 SMR [1.42] (P > 0.05)

Expected based on England and Wales rates; no adjustment for

(1988) London, United

Urinary bladder

1 [0.58] (P > 0.05)

tobacco smoking

Kingdom, 1950–84


153



Soll-Johanning et al. (1998) Denmark, 1943–92

Soll-Johanning et al. (2003) Denmark, 1943–92

15 249 male and 958 female bus and tramway workers

153 cases, 255 controls; all men

Job title obtained from population register; ever employed as an urban bus driver or tramway worker, 1900–94

Job title from population register

Lung

Urinary bladder

Lung

Employed > 3 mo SIR Men* 473 1.6 (1.5–1.8) Men** 390 1.2 (1.1–1.3)

Women* 15 2.6 (1.4–4.3) Men–30 yr since first employed





Exposure categories



Covariates Comments

Petersen et al. P for (2010) trend = 0.79 Denmark, 1979–2003 (cont.)

10-yr lag



Balarajan & McDowall (1988) London, United Kingdom, 1950–84

3392 professional drivers

Job title in 1939 census and still alive in 1950; professional drivers identified from NHS Central Registry

Lung SMR Observed and expected based on England/Wales rates; men only; no adjustment for tobacco smoking

Taxi drivers 30 0.86 (P > 0.05) HGV drivers 280 1.59 (P 0.05) HGV drivers 19 1.06 (P > 0.05)

Gubéran et al. (1992) Geneva, Switzerland, 1949–1986 for mortality, 1970–1986 for incidence

6630 drivers holding a licence in 1949–61

Occupation on driver’s licence

Lung Professional drivers, 15 yr latency

SMR 77 deaths

1.50 (1.23–1.81)

SIR 64 incident cases

1.61 (1.29–1.98)

Mortality (time from first exposure in yr) 0–14 2 0.67 15–24 11 1.18

Expected based on male mortality rates for Switzerland Expected based on mean incident rates for men in Geneva 1970–75, 1976–81, 1982–86

Men born 1900 or later; limited comparison of smoking rates with other men

24–34 24 1.3 35–44 21 1.35 ≥ 45 21 2.59

P for trend = 0.02

Non-professional drivers, 15-yr latency Less exposure group

126 1.21 (1.03–1.40) 97 incident

1.15 (0.97–1.37)

cases More exposure group

24 deaths 24 incident

1.32 (0.91–1.85)

1.61 (1.11–2.27)

cases


157



Gubéran et al. (1992) Geneva, Switzerland, 1949–1986 for mortality, 1970–1986 for incidence (cont.) Hansen (1993) Denmark, 1970–80

Järvholm & Silverman (2003) Sweden, to 1995

Guo et al. (2004a) Finland, records in 1971–95

14 225 HGV drivers and 43 024 unskilled labourers as reference

Swedish construction workers (389 000 total), 1971–92: 6364 male HGV drivers; 110 984 carpenters and electricians as referents

Finnish working population, census data in 1970; 667 121 men

Job in 1970 census

Job recorded at examination

Longest job held in 1970 census

Urinary bladder

Lung

Urinary organs Lung

Lung

Professional drivers, 15-yr latency

HGV drivers

HGV drivers

HGV drivers

HGV drivers

SIR 13 1.25 (0.74–1.99) incident cases

SMR 76 1.60

(1.26–2.00) 11 0.98 (0.49–1.75)

SMR 57 1.37 (1.04–1.78) deaths

SIR 61 1.29 (0.99–1.65)

General population referents

SMR 57 1.18 (0.89–1.53) deaths

SIR 61 1.14 (0.87–1.46)

SIR 620 1.13 (1.04–1.22)

No direct control for tobacco smoking, but similar rates among HGV drivers and referents from population survey

Tobacco smoking (never, current, former, unknown), age at diagnosis or death, calendar time; national industrial health service health examination in 1971–92; smoking history from first examination; general population SIR and SMR not adjusted for smoking Age- and calendar time-adjusted

Indirect adjustment for tobacco smoking

IARC M

ON

OG

RAPH

– 105

158

http:1.04�1.22http:0.87�1.46http:0.89�1.53http:0.99�1.65http:1.04�1.78http:0.49�1.75http:1.26�2.00http:0.74�1.99





Exposure categories



Covariates Comments

Guo et al. (2004b) Finland, 1971–95

Urinary bladder

HGV drivers 144

SIR 1.01 (0.85–1.19)

Same cohort as Guo et al. (2004a); men only; indirect adjustment for tobacco smoking

Laden et al. 54 319, including Job title and industrial Lung SMR Survey of current workers in 2003 (2007) 36 299 union hygiene survey All workers 769 1.04 (0.97–1.12) indicated similar birth cohort-USA, 1985– drivers specific tobacco smoking rates Drivers NR 1.10 (1.02–1.19) 2000 compared with US men; results by Loading dock NR 1.10 (0.94–1.30) job title presented only in figures workers

Urinary All workers 29 0.80 (0.56–1.15) bladder

Garshick et al. (2008) USA, 1985–

31 135 union drivers

Job title and industrial hygiene survey

Lung (underlying or

Change in HR (%) per yr increase

Age, calendar yr, time period of hire, attained age in 1985, time on and off work, race, census region;

2000 secondary cause)

Long-haul drivers Pick-up and delivery drivers

323 233

2.5 (0.2–4.9) 3.6 (1.2–6.1)

men only; indirect adjustment for tobacco smoking using smoking rates from 2003 worker survey

Dock workers 205 3.4 (0.8–6.0) Combination 150 4.0 (1.5–6.6) workers (local drivers/dock workers)

HR for > 1 yr work

Mechanics 38 0.95 (0.66–1.38) Hostlers 29 0.99 (0.68–1.45) Clerks 15 0.55

(0.32–0.95) HR for 20 yr work

Long-haul drivers 323 1.65 (1.04–2.62) 1.40 Smoking-adjusted (0.88–2.24)


159

http:0.56�1.15http:0.94�1.30http:1.02�1.19http:0.97�1.12



Garshick et al. (2008) USA, 1985– 2000 (cont.)

Birdsey et al. 156 241 members None Lung (2010) of US HGV USA, 1989– owners/ operators Bladder 2004 trade association and other

urinary

Garshick et al. 31 135 union Job title and industrial Lung (2012) drivers hygiene survey, USA, 1985– workers employed ≥ 1 2000 yr in 1985

Pick-up and 23 2.04 delivery drivers (1.28–3.25)

2.21 (1.38–3.52) Smoking-adjusted Dock workers 205 1.94 (1.18–3.18)

2.02 (1.23–3.33) Smoking-adjusted Combination 150 2.20 (1.35–3.61) Smoking-adjusted workers (local 2.34 (1.42–3.83) drivers/dock workers) Trade association 557 SMR Expected deaths from US membership 1.00 (0.92–1.09) population; 94% men; only 32% had Trade association 0.93 (0.62–1.34) more than 9 yr of follow-up; overall membership SMR, 0.76, suggesting healthy-

worker effect; no adjustment for tobacco smoking

Cumulative exposure Age, calendar year, time period of (µg/m3–mo) hire, attained age at study entry, No lag race, census region; men only;

mechanics excluded



Adjusted for duration

Garshick et al.



Miners Boffetta et al. (1988) American Cancer

476 648 Self-reported job Lung Miners 15 2.67 (1.63–4.37)

Age and tobacco smoking, including pipe/cigar only; mortality in men aged 40–79 yr in 1982; any work as a miner

Society Cohort, 1982–84 Guo et al. (2004a) Finland, records in 1971–95

Finnish working population in 1970 census; 667 121 men

Longest job held in 1970 census

Lung Metal ore miners

Non-metal ore miners

36

181

SIR 3.26 (2.28–4.51) 1.85 (1.59–2.14)


Other miners 70 1.73 (1.35–2.19) Guo et al. (2004b) Finland, 1971–95

Urinary bladder Non-metal ore

miners 22

SIR 1.16 (0.73–1.76)


Neumeyer-Gromen et al. (2009) Former East Germany, 1970–2001

5862 former East German potash miners employed after 1969

Industrial hygiene job review and total carbon measurements

Lung (mortality) Urinary bladder (mortality) Lung (mortality)

All workers All workers

> 4.9 [mg/m3]–yr

61 8

61

SMR 0.73 (0.57–0.93) 0.80 (0.40–1.60)

1.28 (0.61–2.71)

Follow-up of Säverin et al. (1999); expected based on male population of former East Germany; median duration of exposure, 14.9 yr; mean follow-up, 26 yr since hire

Age, tobacco smoking (nonsmoker, smoker, missing); men; data on smoking available from medical examinations

Exposure quintiles [mg/m3]–yr

Age, smoking, duration of follow-up



P for trend = 0.09

Subcohort with more accurate exposure data > 4.9 [mg/m3]–yr 37 1.50 Age, smoking (nonsmoker, smoker,

(0.66–3.43) missing)

Neumeyer-Gromen et al. (2009) Former East Germany, 1970–2001 (cont.)

> 3.90 15 2.28 (0.87–5.97)

3335

Attfield et al. 12 315 non-metal Modelled estimate SMR SMRs calculated using state(2012) miners in 8 mines of respirable EC Lung Ever underground 122 1.21 (1.01–1.45) specific rates USA, 1947–67 underground; (mortality) Exposures to radon, asbestos and Surface only 81 1.33 (1.06–.66) depending on no historical silica low; no effect on results of Urinary Ever underground 0.69 (0.23–1.62) mine, through measurements for exposure to EC bladder Surface-only 1.68 (0.72–3.30) 1997 surface miners (mortality)

Cumulative EC, 15-yr Lung Ever underground HR Unadjusted for tobacco smoking; lag (mortality) (µg/m3–yr) similarly elevated risks with average



IARC M

ON

OG

RAPH

– 105

Attfield et al. (2012) USA, depending on mine, through 1997 (cont.)

Average EC intensity, 15-yr lag

Cumulative EC, 15-yr lag

Average EC intensity, 15-yr lag

Silverman et al. (2012) USA, depending on mine, through 1997

Non-metal miners in 8 mines, nested case-control study, 1947–67; 198 cases and 562 controls

See Attfield et al. (2012)

Lung

Cumulative EC, no lag

Excluding 57 persons with



Silverman et al. (2012) USA, depending on mine, through 1997 (cont.)

Surface only (µg/m3–yr) OR 0–





Exposure categories



Covariates Comments

Other exposed workers Heavy equipment operators Wong et al. 34 156 members Job category review Lung All members 309 SMR Expected deaths from USA age-, (1985) of a construc-tion 0.99 (0.88–1.10) sex-, cause-specific rates; men only; USA, 1964–1978

union for at least 1 yr

Membership no direct information on tobacco smoking; survey of only 107



Järvholm & Swedish Job recorded at Lung (162) Silverman construction examination (2003) workers (389 000 Sweden, to total), 1971–92: 1995 14 364 operators

of heavy

construction vehicles; 110 984 carpenters and electricians as referents

Guo et al. Finnish working Longest job held on Lung (2004a) population in 1970 1970 census Finland, census; 667 121 records in men 1971–95

Guo et al. Urinary (2004b) bladder Finland, 1971–95

Heavy 49 construction deaths vehicle operators

61

Use of a cabin on construction vehicle Never 10

Sometimes 37

Always 7

Forklift drivers 80

Excavation 76 machine operators Road building 121 machine operators Construction 104 machine operators Forklift drivers 19

Excavation 18 machine operators Road building 23 machine operators Construction 19 machine operators

SMR 0.83 (0.61–1.09)

SIR 0.87 (0.66–1.11) SIR

0.86 (0.4–1.6)

0.71 (0.5–1.0)

0.50 (0.2–1.0) P for trend



Dock workers Gustafsson 6071 Swedish dock None Lung Cohort 70 SMR et al. (1986) workers membership deaths; 1.32 (1.05–1.66) Sweden, 86 SIR 1961–80 incident 1.68

cases (1.36–2.07)

Emmelin et al. Nested case– Duration of use of Lung Yr of work since diesel introduced (1993) Sweden, 1950–74

control study in update of Swedish dock worker cohort: 50 cases/154 controls

diesel equipment and indices related to fuel consumption, 2-yr lag

Low/nonsmoker 9 1.00 Medium 27 1.8 (0.5–6.6) High 14 2.9 (0.6–14.4) Fuel use index Low 10 1.00 Medium 25 1.5 (0.5–4.8) High 15 2.9 (0.7–11.5) Exposure time Low/non-smoker 12 1.00 Medium 19 2.7 (0.6–11.3) High 19 6.8 (1.3–34.9)

Guo et al. Finnish working Longest job held on Lung Stevedores 236 SIR (2004a) population in 1970 1970 census 1.32 (1.16–1.50) Finland, census; 667 121 records in men 1971–95 Guo et al. Urinary Stevedores 31 SIR (2004b) bladder 0.95 (0.65–1.35) Finland, 1971–95

Lung cancer deaths and cases linked to national cancer and mortality register; expected rates from county/metro area of workers; diesel HGVs first in ports in the late 1950s, increased use in the 1960s; no information on tobacco smoking Lung cancer deaths and cases linked to national cancer and mortality register; adjusted for tobacco smoking and exposure variables; controls matched on port and date of birth; results presented with 90% CI



IARC M

ON

OG

RAPH

– 105

168





Exposure categories



Covariates Comments

Occupations entailing exposure to diesel exhaust Boffetta et al. (1988) American Cancer

476 648 Self-reported exposure to diesel exhaust at work in 1982

Lung Any exposure 1–15 yr ≥ 16 yr

174 1.18 (0.97–1.44) 1.05 (0.80–1.39) 1.21 (0.94–1.56)

Age strata, tobacco smoking, other occupational exposures; mortality in men aged 40–79 yr in 1982

Society Cohort, 1982–84 Van Den 160 230 Questionnaire; self- Lung Exposure in past 1 662 1.13 (0.93–1.36) Age, gender, education, race and Eeden & reported exposure in yr tobacco smoking; health plan Friedman (1993) California, USA, 1964–88

past and previous to past yr

Urinary bladder

Exposure in past yr and before Exposure in past yr Exposure in past

650

1.02 (0.81–1.29)

1.17 (0.86–1.59)

1.16 (0.81–1.67)

participants, aged 18–79 yr, free of cancer at time of examination

yr and before Boffetta et al. Men, 28 million Linkage to Swedish Lung Any diesel 6 266 SIR Expected based on Swedish rates; (2001) PY; women, 15 Cancer Environment exposure 1.09 (1.06–1.12) age, calendar period, region and Sweden, 1971–89

million PY Register for job and industry title from 1960 census; JEM

Unexposed Low probability

17 979 2 222

1.0 (reference) 1.1 (1.04–1.13)

urban/rural residence; rates for men shown; few cases in women; no information on tobacco smoking

Medium 1 881 0.90 probability (0.86–0.94) High probability 1 841 1.2 (1.10–1.21)

Urinary Unexposed 12 287 1.0 (reference) bladder Low probability 1 380 0.99 (0.94–1.05)

Medium 1 220 0.84 probability (0.79–0.89) High probability 1 069 0.98 (0.92–1.04)

Zeegers et al. 58 279 from JEM from job history Urinary No exposure 428 1.0 Age, other occupational exposures, (2001) 204 municipal bladder Low 35 1.00 intensity and duration of cigarette Netherlands, 1986–92

population registries Medium 31

(0.65–1.54) 0.96 (0.60–1.53)

smoking

High 32 1.17 (0.74–1.84)


169

170 Table 2.1 (continued)


Guo et al. (2004a) Finland, records in 1971–95

Finnish working population in 1970 census; 667 121 men

JEM using longest held job in 1970 census, 20-yr lag

Lung None 26 723 1.0 (reference) Asbestos, quartz dust, SES, age, time period; indirect adjustment for tobacco smoking

Lowest 2 436 0.98 (0.94–1.03) Middle 758 1.04 (0.97–1.12) Highest 120 0.95 (0.83–1.10)

Guo et al. JEM using longest Urinary None 6 026 1.0 (reference) Same cohort as Guo et al. (2004a); (2004b) Finland, 1971–95

held job in 1970 census, 20-yr lag

bladder None Lowest Middle

5 872 493 200

1.0 (reference) 1.00 (0.91–1.11) 0.95 (0.83–1.10)

adjusted for tobacco smoking, SES, age and time period; men only

Highest 78 0.97 (0.77–1.21) CI, confidence interval; EC, elemental carbon; HR, hazard ratio; ICD, International Classification of Diseases; IRR, incidence rate ratio; JEM, job–exposure matrix; mo, months; NHS, national health service; NR, not reported; OR, odds ratio; PY, person–years; SES, socioeconomic status; SIR, standardized incidence ratio; SMR, standardized mortality ratio; yr, year

IARC M

ON

OG

RAPH

– 105


few for a meaningful analysis. [Although exposure to coal dust has not been proven to be associated with the risk for lung cancer, it is probable that workers who incurred this exposure also had concurrent exposure to coal combustion before conversion to diesel-powered locomotives. Therefore, although these results are consistent with an increased risk of lung cancer attributable to exposure to diesel exhaust, it is possible that coal combustion products contributed to this risk.]

Boffetta et al. (1988) examined the relationship between lung cancer and occupational exposure to diesel exhaust using data from a prospective mortality study including more than 1.2 million American men and women that was begun in 1982 by the American Cancer Society. Volunteers from across the USA were enrolled by completing a questionnaire that included information on tobacco smoking, current and last jobs, and the job held for the longest period, other exposures and self-reported exposure to diesel exhaust. An assessment of mortality up to September 1984 was conducted in men aged 40–79 years at enrolment. Of these, 2973 men reported working in the railroad industry during any period of their life. Their relative risk for lung cancer was 1.59 (95% confidence interval [CI], 0.94–2.69; based on 14 deaths), compared with men who did not report working in the railroad industry and who reported no exposure to diesel exhaust, after adjusting for age and smoking. Similar results were obtained for those who reported railroad work as their principal occupation. The risk for lung cancer among railroad workers who reported exposure to diesel exhaust was not stated (approximately 50% of persons who had railroad work as their principal occupation reported exposure to diesel exhaust). [The Working Group found some indication of an increased risk of lung cancer, although this was weakened by the use of self-reporting to assess exposure.]

Nokso-Koivisto & Pukkala (1994) studied 8391 members of the Finnish Locomotive Drivers’ Association and determined their cancer incidence between 1953 and 1991 by linkage to the Finnish Cancer Registry. The standardized incidence ratio (SIR) was 0.86 (95% CI, 0.75–0.97; 236 cases) for lung cancer, 1.08 (95% CI, 0.80–1.43; 48 cases) for urinary bladder cancer, 1.25 (95% CI, 0.88–1.17) for kidney cancer and 1.12 (95% CI, 0.92–1.32) for prostate cancer. [These locomotive drivers had also been exposed to asbestos while in training, to coal combustion products (in the 1950s) and to diesel exhaust thereafter. However, in the Finnish railroad industry, much overlap occurred between the periods of steam, diesel and electric locomotive use. Because of the lack of specific information that linked job titles and duties to periods of diesel exhaust use, and in the absence of an internal comparison group, this study was regarded as uninformative in relation to associations between cancer and exposure to diesel exhaust.]

Garshick et al. (2004) studied mortality from lung cancer in 54 973 railroad workers in the USA between 1959 and 1996 (38 years). This study was an update of an earlier study of the same cohort (Garshick et al., 1988), and a pilot study by Schenker et al. (1984). The cohort comprised a sample of men aged 40–64 years and with 10–20 years of railroad service in 1959. Work histories and death follow-up were extended to 1996. The sample comprised approximately 75% of subjects in diesel exhaust-exposed jobs (engineers, conductors, brakemen, hostlers and shop workers) and 25% in jobs with low or no exposure (ticket agents, clerks and signal maintainers), determined from job categories in 1959. Exposure assignment was validated by an industrial hygiene review of current and historical jobs and work practices and measurement of exposure in current workers (Woskie et al., 1988a, b). By 1959, the railroad industry in the USA had largely converted from coal-fired to diesel-powered locomotives and, in this

171

http:0.92�1.32http:0.88�1.17http:0.80�1.43http:0.75�0.97http:0.94�2.69


study, exposure to diesel exhaust was considered to have begun in 1959. Work histories were obtained from the US Railroad Retirement Board, and mortality was ascertained using Railroad Retirement Board, Social Security and Health Care Financing Administration records. Cause of death was obtained from the National Death Index and death certificates. A total of 43 593 deaths occurred, including 4351 from lung cancer. Analyses consisted of internal comparisons (using workers with low or no exposure as the referents). Efforts were made to adjust for a healthy-worker survivor effect by including a variable in the models for total years worked and also terms for time off work.

Compared with workers with low and no exposure, exposed workers had a relative risk for mortality from lung cancer of 1.40 (95% CI, 1.30–1.51), using a 5-year lag, which did not increase with increasing number of years worked in these jobs. The increased risk was associated with all groups stratified by age in 1959, with the exception of the oldest group (aged 60–64 years). Indirect adjustment for tobacco smoking using the methods of Schlesselman and Axelson (Schlesselman, 1978; Axelson & Steenland, 1988), based on job-specific smoking information from a survey among 547 railroad workers in 1982 and an accompanying case–control study (Garshick et al., 1987; Larkin et al., 2000), suggested some positive confounding which would account for a decrease of about 10–20% in exposed versus unexposed railroad workers. Adjustment using these estimated effects of smoking resulted in rate ratios of 1.17–1.27 for exposed versus unexposed railroad workers. [The Working Group noted that mechanics did not have an elevated risk of lung cancer, probably due to exposure misclassification, because this job title included workers with and without repair shop-related exposures to diesel exhaust.]

An additional approach to adjust for cigarette smoking (Garshick et al., 2006) used information on age- and job-specific cigarette smoking

histories available from a previous case–control study of railroad workers based on US Railroad Retirement Board records (Garshick et al., 1987). Data on cause of death, birth cohort, and age- and job-specific smoking habits were used to simulate the smoking behaviour of 39 388 deceased railroad workers. Unadjusted for smoking, the risk of lung cancer among exposed workers with a 5-year lag was 1.35 (95% CI, 1.24–1.46), and an excess risk remained after adjustment (1.22; 95% CI, 1.12–1.32).

To improve the estimation of historical exposures during the transition from steam to diesel locomotives (starting in 1945 and during the decade of the 1950s), yearly locomotive rosters from builders’ records and company-specific information were obtained starting in 1945 (Laden et al., 2006). Although only information on the last railroad worked was available in the computerized database, a review of a sample of work history records of railroad employers that were available on paper indicated that the majority of workers (95%) did not change railroads during their careers. The rosters were used to determine the make, model and horse power (hp) of each locomotive in service for 93% of the eligible cohort (52 812 subjects). From this information, the number and type of locomotives that were diesel-fuelled were calculated annually for each railroad. An estimate of the relative amount of PM produced by diesel locomotives (in grams of particulate per hour) for a given railroad in a given year was estimated using Environmental Protection Agency emission factors and information on engine horse power. The probability of diesel exposure per year per railroad (diesel fraction) was then calculated using the year-specific ratios of grams of particulate until this value became constant or the railroad was known to be 100% diesel-powered. Years of exposure to diesel exhaust were calculated by weighting yearly months worked with a specific railroad and the yearly diesel fraction, which allowed the estimation of years of exposure for each worker

172

http:1.12�1.32http:1.24�1.46http:1.17�1.27http:1.30�1.51


before 1959 in contrast to the previous analysis (Garshick et al., 2004), in which exposure was considered to start in 1959. Among workers hired in 1945–49, who started work when diesel locomotives were introduced into the industry, the relative risk of lung cancer with a 5-year lag for any exposure was 1.77 (95% CI, 1.50–2.09) and, for workers hired in 1939–44, was 1.30 (95% CI, 1.19–1.43; P for interaction = 0.003). There was evidence of an exposure–response relationship with years of exposure duration that plateaued at 15 years and was not present for workers hired before 1945. Compared with no exposure, the relative risks for exposure for 0–


2.2.2 Bus garage workers

Gustavsson et al. (1990) studied mortality from, and incident cases of, lung cancer in a cohort of 695 Stockholm (Sweden) bus mechanics, servicemen and hostlers who had worked for at least 6 months from 1945 to 1970. Mortality was assessed from 1952 to 1986 and incidence from 1958 to 1984. Diesel-powered buses were first introduced in the 1930s in Stockholm and all buses were diesel-fuelled after 1945. A JEM was designed by industrial hygienists to categorize the intensity of exposure to diesel exhaust and asbestos by work period and specific workplace. Specific exposure measurements were limited, and relative exposures were therefore estimated on the basis of work practice, number and characteristics of buses, and garage ventilation. Intensity was estimated over six levels, starting at zero, and cumulative exposure was calculated (intensity × years) for each period and summed. A nested case–control study was performed to match each incident lung cancer case (n = 20) to six controls by age. Relative to the lowest category of cumulative exposure to diesel exhaust, an increased risk was observed with increasing categories of the exposure index: 10–20–relative risk, 1.27 (95% CI, 0.21–7.72; two cases); 20–30– relative risk, 1.56 (95% CI, 0.34–7.16; three cases); and > 30–relative risk, 2.63 (95% CI, 0.74–9.42; 10 cases). The relative risk per unit of a continuous diesel exhaust ‘score’ was 1.37 (95% CI, 0.91–2.07). No evidence of an exposure–response relationship with asbestos score was observed. Compared with mortality rates of other occupationally active men, the standardized mortality ratio was 1.22 (95% CI, 0.71–1.96; 17 deaths) for lung cancer and 1.23 (95% CI, 0.45–2.68; six deaths) for haematopoietic cancer. Compared with the Stockholm general population, the standardized mortality ratio was 1.93 (95% CI, 0.53–4.94; four deaths) for oesophageal cancer, 0.97 (95% CI, 0.32–2.27; five deaths) for stomach cancer and 0.52 (95% CI, 0.01–2.88; one death)

for urinary bladder cancer. [The Working Group noted that no information on tobacco smoking was available. However, confounding by smoking based on diesel exposure category was improbable. The major limitation of this study was the small number of cases, which limited statistical inferences.]

2.2.3 Professional drivers

(a) Bus drivers

Balarajan & McDowall (1988) used the National Health Service Central Register to identify 3392 men, who were employed as professional drivers, and were required to hold a professional licence, in London (United Kingdom) according to the 1939 census, and were alive in January 1950. Mortality compared with the general population was subsequently assessed from 1 January 1950 until the end of 1984. Among bus drivers, the risk for lung cancer (SMR, 1.42 [95% CI, 0.84–2.24]; 18 deaths) was not significantly elevated. [The Working Group calculated the exact 95% confidence intervals that were not provided by the authors.] No significant increase in risk of mortality was observed for cancer of the stomach (SMR, 1.68; 95% CI, 0.72–3.31; eight deaths) or for urinary bladder cancer (SMR, 0.58; 95% CI, 0.015–3.23; one death). No deaths from leukaemia or other lymphatic neoplasms occurred. A significantly elevated standardized mortality ratio was found for bronchitis, emphysema and asthma (SMR, 1.66; 95% CI, 1.06–2.47; 24 deaths). [The Working Group noted that no information was available regarding tobacco smoking, and the relationship between occupational title and the specific periods of work with exposure to diesel exhaust was not described. These drivers were unlikely to have been exposed to appreciable levels of diesel exhaust before the 1950s, when diesel engines were introduced.]

Soll-Johanning et al. (1998) conducted a retrospective cohort study of 18 174 bus drivers and tramway employees in Copenhagen (Denmark)

174

http:1.06�2.47http:0.015�3.23http:0.72�3.31http:0.84�2.24http:0.01�2.88http:0.32�2.27http:0.53�4.94http:0.45�2.68http:0.71�1.96http:0.91�2.07http:0.74�9.42http:0.34�7.16http:0.21�7.72


who were employed 1900–94. In Copenhagen, the first diesel-powered buses were introduced in 1936, but, during the Second World War, all buses were fuelled with gasoline. Diesel-powered buses gradually replaced gasoline-powered models after that time, and, in the 1960s, they also replaced the trams. Cancer rates were compared with those of the general population of Denmark by linkage to the Danish Cancer Registry and National Death Index to identify cancers that occurred after 1943. Among male workers employed for 3 months or longer, the standardized incidence ratio was 1.6 (95% CI, 1.5–1.8; 473 cases) for lung cancer, 1.0 (95% CI, 0.8–1.3; 82 cases) for stomach cancer, 1.2 (95% CI, 1.0–1.5; 105 cases) for rectal cancer, 1.4 (95% CI, 1.0–1.9; 39 cases) for laryngeal cancer, 1.6 (95% CI, 1.3–1.6; 83 cases) for kidney cancer, 1.4 (95% CI, 1.2–1.6; 177 cases) for urinary bladder cancer, 1.9 (95% CI, 1.2–2.8; 22 cases) for pharyngeal cancer and 1.1 (95% CI, 0.8–1.5; 46 cases) for leukaemia. In women, the standardized incidence ratio for lung cancer was 2.6 (95% CI, 1.5–4.3; 15 cases). In both men and women, a greater risk of lung cancer was observed with longer time since first employment. No trend in lung cancer risk was found based on periods of predominantly gasoline or diesel vehicle use, and the risks were similarly elevated for workers who started before, at the onset, or during the use of diesel buses. [The Working Group noted that no information on specific exposures or tobacco smoking was available and the periods of diesel and gasoline emissions overlapped. Compared with other men in Copenhagen, the smoking rates among the bus drivers were slightly higher during some time periods, suggesting the possibility of some confounding by smoking.]

A nested case–control study (Soll-Johanning et al., 2003) was conducted with 153 cases of lung cancer and 84 cases of urinary bladder cancer included in the cohort of Copenhagen bus drivers and tramway employees. The cases and controls or next of kin were interviewed

regarding tobacco smoking history. Deaths from cancer or non-neoplastic respiratory disease were excluded from the control group and cases and controls were matched on date of birth. Both 10-year lag and no lag models, based on duration of employment, were assessed, adjusting for smoking history in seven categories based on pack–years. No consistent increase in lung cancer risk was observed based on categories of duration of employment in either lag model. The risk, although not statistically significant, increased with greater number of years of employment, but then decreased after > 20 years. With a 10-year exposure lag, there was a suggestion of an increased risk for urinary bladder cancer in persons with 10–


history, information regarding bus routes and tobacco smoking habits. Information on incident cases of cancer up to 2003 was obtained by linkage to the Danish Cancer Registry. In analyses comparing external rates for men in the three cities, the standardized incidence ratios were 1.2 (95% CI, 1.0–1.4; 100 cases) for lung cancer and 1.6 (95% CI, 1.2–2.0; 69 cases) for urinary bladder cancer; no cancer at other sites had an elevated risk. The risk for urinary bladder cancer was increased for drivers employed for 15 years or longer (SIR, 1.81; 95% CI, 1.2–2.5) and for drivers with less than 15 years of employment (SIR, 1.5; 95% CI, 1.0–2.1). The standardized incidence ratio for lung cancer was also marginally increased for bus drivers with 15 years or more of employment (SIR, 1.3; 95% CI, 1.0–1.8). A Cox regression model was used to assess the relationship between risk and duration of employment. After adjustment for smoking, city of employment and usual type of bus route operated (urban versus rural), in addition to age and calendar time, each additional year of employment as a bus driver was associated with slightly, non-significantly increased risks for bladder cancer (RR, 1.02; 95% CI, 0.99–1.05). No overall increased risk was observed for lung cancer (RR, 1.00; 95% CI, 0.98–1.03). In a comparison of drivers employed for 15–24 and 25 years or longer with those with those employed for less than 15 years, the relative risks were 1.1 and 1.3, respectively, for bladder cancer and 0.9 and 1.0, respectively, for lung cancer. No change in the estimates was found in a 10-year lag model. [The Working Group noted that these data indicated that, when adjusted for smoking and other risk factors and using an internal comparison group, there was little to no increased risk for urinary bladder or lung cancer in bus drivers with increasing duration of employment. This finding contrasted with the elevated risks for bus drivers suggested by the standardized incidence ratio results reported. Adjustment for the type of bus route (urban versus rural) may

have limited the ability to demonstrate an effect of occupational exposure to diesel exhaust.]

(b) Heavy goods vehicle and other drivers

In the USA, Menck & Henderson (1976) reviewed 2161 death certificates that reported lung cancer (trachea, bronchus and lung) in white men, aged 20–64 years, in 1968–70 and all 1777 incident cases of lung cancer in white men of the same age reported to the Los Angeles County Cancer Surveillance Program in 1972– 73. These mortality and morbidity data were pooled because of the high accuracy of lung cancer death ascertainment and high mortality. Information on either occupation or industry was not available for 1911 subjects. The population at risk by age group, occupation and industry was estimated from a ‘1-in-50’ sample of Los Angeles County white men, aged 20–64 years, obtained from the 1970 census for Los Angeles. Expected deaths and expected incident cases were calculated for each specific occupation, assuming that the age-specific rates of cancer in each occupation were the same as those for all occupations. The standardized mortality ratio was 3.44 ([95% CI, 2.18–5.16]; 16 deaths, seven incident cases) among taxi drivers and 1.65 ([95% CI, 1.35–1.99]; 58 deaths, 51 incident cases) among heavy goods vehicle (HGV) drivers. [The Working Group calculated the exact 95% confidence intervals, which were not provided by the authors. This study was limited by the absence of data and the use of the occupation recorded on the death certificate as a proxy for exposure.]

In the study by Boffetta et al. (1988) described in Section 2.2.1, 9738 men stated their main occupation as an HGV driver, of whom 47% reported exposure to diesel exhaust, 33% reported no such exposure and 20% did not respond. Among the 16 208 HGV drivers, based on any past employment in this occupation, the relative risk for lung cancer was 1.24 (95% CI, 0.93–1.66; 48 deaths) compared with other men who were not HGV drivers and who reported no employment in a

176

http:0.93�1.66http:1.35�1.99http:2.18�5.16http:0.98�1.03http:0.99�1.05


job that entailed exposure to diesel exhaust, after controlling for age and tobacco smoking. No difference in relative risks for lung cancer was observed between HGV drivers who reported exposure to diesel (RR, 1.22; 95% CI, 0.77–1.95; 18 deaths) and those who reported no exposure to diesel (RR, 1.19; 95% CI, 0.74–1.89; 18 deaths). However, in a direct comparison between drivers exposed and those not exposed to diesel exhaust, a suggestion of a positive trend with duration was found for the diesel-exposed HGV drivers (duration 1–15 years–RR, 0.87; 95% CI, 0.33–2.25; six exposed deaths; duration > 16 years–RR, 1.33; 95% CI, 0.64–2.75; 12 exposed deaths). [The Working Group noted that this study was limited by the small number of HGV drivers who reported exposure, and categorization of exposure by self-reporting.]

In the study by Balarajan & McDowall (1988) described in Section 2.2.2, a significant excess of lung cancer was observed among HGV drivers (SMR, 1.59 [95% CI, 1.41–1.79]; 280 deaths), but not among taxi drivers (SMR, 0.86 [95% CI, 0.58–1.23]; 30 deaths). [The Working Group calculated the exact 95% confidence intervals which were not provided by the authors.] Among HGV drivers, the standardized mortality ratio for cancer of the stomach was 1.41 ([95% CI, 1.11–1.77]; 73 deaths), but was not significantly elevated for urinary bladder cancer (SMR, 1.06; 19 deaths), leukaemia (SMR, 1.02; nine deaths) or other lymphatic neoplasms (SMR, 0.92; 12 deaths). Among taxi drivers, no significant increase in risk of mortality was found for cancer of the stomach (SMR, 0.68; eight deaths), urinary bladder cancer (SMR, 1.21; five deaths), leukaemia (SMR, 1.64; three deaths) or other lymphatic neoplasms (SMR, 1.61; four deaths). [The Working Group noted that the interpretation of this study was limited by the lack of information on tobacco smoking.]

Rafnsson & Gunnarsdóttir (1991) identified 868 HGV drivers and 726 taxi drivers in Reykjavik (Iceland) in 1951 from union records

and assessed mortality up to 1988. The standardized mortality ratio was 2.14 (95% CI, 1.37–3.18; 24 deaths) for HGV drivers and 1.39 (95% CI, 0.72–2.43; 12 deaths) for taxi drivers. The risk for HGV drivers was also evaluated on the basis of duration of employment and no variation was found.

Gubéran et al. (1992) conducted a cohort mortality and cancer morbidity study of 6630 drivers from the Canton of Geneva (Switzerland) who were licensed from 1949 to 1961. The exposed group was compared with the general population of Geneva. The drivers were distributed into three groups: (1) professional drivers (n = 1726), (2) non-professional drivers ‘more exposed’ to exhaust gas and fumes (this group included occupations such as vehicle mechanic, policeman and road sweeper; n = 712), and (3) non-professional drivers ‘less exposed’ (comprising all other occupations; n = 4192). The cohort was followed up from 1949 to 1986 for mortality and from 1970 to 1986 for cancer morbidity. With a 15-year latency, significant excesses for lung cancer mortality (SMR, 1.50; 90% CI, 1.23–1.81; 77 deaths) and morbidity (SIR, 1.61; 90% CI, 1.29–1.98; 64 cases) were observed among professional drivers, and the risk of lung cancer increased significantly with time from first exposure. Among non-professional drivers, no significant excess risk was found except for lung cancer mortality among the ‘less exposed’ group (SMR, 1.21; 90% CI, 1.03–1.40), and for the incidence of lung cancer among the ‘more exposed’ group (SIR 1.61; 90% CI, 1.11–2.27). For 15 years of latency, the authors reported a standardized mortality ratio of 1.43 (95% CI, 0.80–2.36; 11 deaths) and a standardized incidence ratio of 1.25 (95% CI, 0.74–1.99; 13 cases) for urinary bladder cancer. An excess of mortality and morbidity from rectal cancer (SMR, 2.58; 95% CI, 1.62–3.92, 16 deaths; SIR, 2.00; 95% CI, 1.27–3.00; 17 cases) and stomach cancer (SMR, 1.79; 95% CI, 1.17–2.63; 19 deaths; SIR 2.33; 95% CI, 1.56–3.36; 21 cases) was also observed. No significantly increased risk was

177

http:1.56�3.36http:1.17�2.63http:1.27�3.00http:1.62�3.92http:0.74�1.99http:0.80�2.36http:1.11�2.27http:1.03�1.40http:1.29�1.98http:1.23�1.81http:0.72�2.43http:1.37�3.18http:1.11�1.77http:0.58�1.23http:1.41�1.79http:0.64�2.75http:0.33�2.25http:0.74�1.89http:0.77�1.95


found for leukaemia or lymphatic cancers, prostate cancer or cancer at other sites. [The Working Group noted that this study was limited by a lack of specific information on tobacco smoking, and uncertainty on the extent to which these broad occupational titles indicated exposure to diesel exhaust.]

Hansen (1993) identified a cohort of 14 225 HGV drivers and 43 024 other workers based on occupations reported in the 1970 Danish census. The group of non-HGV drivers included unskilled labourers from other industries with no exposure to combustion products. Up to 1980, 627 drivers and 3811 non-drivers died. The risk for mortality from lung cancer was significantly elevated (SMR, 1.60; 95% CI, 1.26–2.00; 76 cases), whereas mortality from urinary bladder cancer (SMR, 0.87; 95% CI, 0.32–1.89) and cancer of the blood and lymph-forming tissues (ICD-8 200–209; SMR, 1.26; 95% CI, 0.78–1.92; 21 deaths) was not increased. [The Working Group noted that the interpretation of these results was limited by the lack of information on specific exposures and tobacco smoking. However, the use of a blue-collar, non-HGV driver comparison group was liable to reduce possible confounding by smoking.]

Järvholm & Silverman (2003) analysed a computerized register of Swedish construction workers (389 000 workers) who participated in health examinations to assemble a cohort of male HGV drivers (n = 6364) and drivers of heavy construction vehicles (see Section 2.2.5). Carpenters/electricians constituted the reference group (n = 119 984). Workers were identified from health examinations in 1971–92 and were linked to the Swedish National Cancer Registry and National Death Registry, from which cases of lung cancer were ascertained up to 1995. For the analysis, data were stratified into never, former and current smokers, on the basis of tobacco smoking habits recorded at the baseline or when available. HGV drivers had an increased risk of lung cancer after adjusting for smoking

(SIR, 1.29; 95% CI, 0.99–1.65; 61 cases); it was not possible to conduct an analysis based on duration and latency. In the same cohort, a significant excess of prostate cancer was also found among HGV drivers (SIR, 1.24; [95% CI, 1.04–1.48]; 124 cases). No significant excess of cases of cancer at other sites was observed, including laryngeal cancer (SIR, 1.25; seven cases), urinary tract cancer [probably bladder] (SIR, 0.72; 26 cases), nasopharyngeal cancer (SIR, 0.82; 12 cases), stomach cancer (SIR, 1.23; 27 cases), rectal cancer (SIR, 1.46; 35 cases), kidney cancer (SIR, 1.12; 23 cases), and lymphoma and leukaemia (SIR, 1.21; 53 cases). [The Working Group noted that the interpretation was limited by the lack of specific information on exposure to diesel exhaust. The strengths of the study were the adjustment for smoking and the use of another blue-collar comparison group, which was liable to reduce possible confounding.]

In the study by Guo et al. (2004a) described in Section 2.2.1, among male HGV drivers who were exposed to both gasoline and diesel exhaust, the risk of lung cancer was elevated (SIR, 1.13; 95% CI, 1.04–1.22; 620 cases); the standardized incidence ratio for taxi drivers was 1.10 (95% CI, 0.96–1.26; 209 cases). In a separate report, Guo et al. (2004b) presented data for other cancers among male HGV drivers. The risk of leukaemia (SIR, 1.29; 95% CI, 1.02–1.60; 82 cases) was significantly elevated, but not that of kidney cancer (SIR, 1.00; 95% CI, 0.84–1.19; 131 cases), oesophageal cancer (SIR, 1.10; 95% CI, 0.76–1.54; 34 cases) or urinary bladder cancer (SIR, 1.01; 95% CI, 0.85–1.19; 144 cases). In male taxi drivers, the risk of kidney cancer was significantly elevated (SIR, 1.39; 95% CI, 1.06–1.79; 61 cases), but not that of oesophageal cancer (SIR, 1.28; 95% CI, 0.70–2.15; 14 cases), urinary bladder cancer (SIR, 1.06; 95% CI, 0.80–1.38; 55 cases) or leukaemia (SIR, 1.09; 95% CI, 0.70–1.62; 24 cases). [The Working Group noted that these studies were limited by the lack of detailed work histories.]

178

http:0.70�1.62http:0.80�1.38http:0.70�2.15http:1.06�1.79http:0.85�1.19http:0.76�1.54http:0.84�1.19http:1.02�1.60http:0.96�1.26http:1.04�1.22http:1.04�1.48http:0.99�1.65http:0.78�1.92http:0.32�1.89http:1.26�2.00


A cohort was assembled from the records of four transport companies in the USA, from which 54 319 male workers employed in 1985 were identified (Laden et al., 2007). Cause-specific mortality was assessed up to 2000 using the National Death Index; 769 deaths from lung cancer were ascertained. Standardized mortality ratios and 95% confidence intervals were calculated for the entire cohort and by job title, using US mortality rates as the referent. Rates of lung cancer were elevated among all drivers (SMR, 1.10; 95% CI, 1.02–1.19) and loading dock workers (SMR, 1.10; 95% CI, 0.94–1.30). The standardized mortality ratio for urinary bladder cancer was 0.80 (95% CI, 0.56–1.15; 29 deaths) for the entire cohort, with similar estimates for drivers and non-drivers.

Garshick et al. (2008) conducted an internal analysis in the same cohort, based on employment records for 31 135 male workers, aged 40 years and over, with at least 1 year of employment as of 1985. Exposure to engine exhaust was estimated for eight job categories (long haul driver, pick-up and delivery driver, loading dock worker, combined driver and dock worker, mechanic, hostler in the vehicle yard, clerk and other) through an industrial hygiene review of previous exposures and current measurements of work-shift exposures to EC. Time-varying cumulative years of work in each of the categories were calculated. Group level adjustment for cigarette smoking was carried out using the methods of Schlesselman and Axelson (Schlesselman, 1978; Axelson & Steenland, 1988). Job-specific information on the distribution of smoking habits was obtained from a survey of 11 986 workers that included all clerks and a random sample of active and retired workers from three of transport companies who contributed to the cohort. In the assessment of mortality from lung cancer, all eight job-specific exposure variables were included in Cox regression models to adjust the risk for lung cancer for different jobs held throughout a worker’s career. The healthy-worker survivor effect

was controlled for using variables for duration of employment and time since leaving work. Hazard ratios (HRs) for lung cancer were elevated in workers who held jobs associated with regular exposure to vehicle exhaust, including long-haul drivers (HR, 1.15, 95% CI, 0.92–1.43; 323 deaths), pick-up and delivery drivers (HR, 1.19; 95% CI, 0.99–1.42; 233 deaths), dock workers (HR, 1.30; 95% CI, 1.07–1.58; 205 deaths) and combination dock workers/drivers (HR, 1.40; 95% CI, 1.12–1.73; 150 deaths). No excess risk was seen in the other job groups. The risk of mortality increased linearly with years of employment in long-haul drivers, pick-up and delivery, dock workers and combination dock workers/drivers. Estimates of risk for 20 years of employment in a job versus no employment in the job, adjusted for smoking, were 1.40 (95% CI, 0.88–2.24) for long-haul drivers, 2.21 (95% CI, 1.38–3.52) for pick-up and delivery drivers, 2.20 (95% CI, 1.23–3.33), for dock workers and 2.34 (95% CI, 1.42–3.83) for combination drivers/dock workers. Risks not adjusted for smoking were slightly greater than the adjusted estimates for long-haul drivers and slightly lower for pick-up and delivery drivers and for dock workers. [The Working Group noted that this study had particular strengths because detailed historical work records were available, exposures in each job were supported by an industrial hygiene review and exposure measurements, and variation in smoking behaviour by job was considered. Compared with the standardized mortality ratio analysis of the same cohort by Laden et al. (2007), greater risks were observed in this study that used an exposure assessment and an internal comparison of risks based on job titles. Although the risks for lung cancer in mechanics, who had greater current and historical exposures, was not elevated, it was noted that the number of mechanics was relatively small (6% of the cohort) and contributed relatively few cases of lung cancer (n = 38). The Working Group also noted that there was no consensus on the optimal method to adjust for

179

http:1.42�3.83http:1.23�3.33http:1.38�3.52http:0.88�2.24http:1.12�1.73http:1.07�1.58http:0.99�1.42http:0.92�1.43http:0.56�1.15http:0.94�1.30http:1.02�1.19


the healthy-worker survivor bias, and that the applied adjustment probably did not negate the positive findings.]

Birdsey et al. (2010) conducted a cohort mortality study of independent HGV owners/ operators using files from a trade association. The 156 241 subjects were members of the trade association between 1989 and 2004, and mortality was assessed using the National Death Index up to 2004. Indirect adjustment was made for tobacco smoking. Most of the cohort (86%) was aged 25–54 years at entry into the study. No excess mortality from lung cancer (SMR, 1.00; 95% CI, 0.92–1.09; 557 deaths), urinary bladder cancer (SMR, 0.93; 95% CI, 0.62–1.34; 29 deaths) or any other type of cancer was observed. [The Working Group noted that the interpretation of this study was limited by the lack of individual information on smoking and information on how association membership served as a surrogate for exposures to exhaust.]

Garshick et al. (2012) investigated the risk for lung cancer in the cohort of Laden et al. (2007) in relation to a reconstruction of occupational exposure to EC. An exposure assessment was conducted in 2001–06 by collecting more than 4000 cross-shift samples of EC measured in PM ≤ 1.0 µm diameter at representative transport terminals. Separate exposure models were constructed for drivers and terminal workers. Historical trends in the ambient levels of EC in terminals were modelled on the basis of historical trends available for 1971–2000 in the coefficient of haze, a measurement of PM based on optical density that is highly predictive of ambient EC. A 1988–89 assessment of EC in the same industry that used the same methodology as the current assessment was used to calibrate model estimates. Historical data on jobs and terminal-specific monthly EC concentrations determined by the exposure model were summed by year for 1971–2000 to estimate cumulative exposure (µg/m3–months) for individuals. Job-specific EC values before 1971 (8% of total exposure time)

were assigned values for 1971 exposures because data on coefficient of haze were not available to estimate background. Combination workers were assumed to spend 50% of their time as a pick-up and delivery driver and 50% as a dock worker. When adjusted for race, calendar year and census region, with a 5-year lag, mortality from lung cancer was elevated for the upper three cumulative EC quartiles compared with the lowest quartile, but the differences were not statistically significant (HR, 1.17–1.19 for 5-year lagged exposures excluding mechanics). However, the risk for lung cancer was inversely associated with the total duration of employment. The association of lung cancer with cumulative exposure to EC was stronger after adjustment for duration of employment, and when mechanics were excluded. The job duties of mechanics changed over time and their exposures were intermittent. The risks for 5- and 10-year lagged exposures increased with each cumulative exposure quartile when mechanics were excluded, resulting in estimated hazard ratios of 1.48 (95% CI, 1.05–2.10) and 1.41 (95% CI, 0.95–2.11) for the highest versus lowest quartiles of 5- and 10-year lagged exposures, respectively, when adjusted for duration of employment. Associations were weaker for average exposure to EC. In addition, adjusting for duration of employment, a linear exposure–response relationship was suggested when cumulative EC was used as a continuous covariate and splines were incorporated into the models. For each 1000 µg/m3–months of cumulative EC, based on a 5-year exposure lag, the hazard ratio was 1.07 (95% CI, 0.99–1.15) with a similar association for a 10-year exposure lag (HR, 1.09; 95% CI, 0.99–1.20). [The Working Group noted that this analysis provided similar results to those of Garshick et al. (2008), who did not use quantitative exposure measurements. An additional strength was a comprehensive exposure assessment and the development of exposure models that were linked to accurate historical job title records and incorporated

180

http:0.99�1.20http:0.99�1.15http:0.95�2.11http:1.05�2.10http:1.17�1.19http:0.62�1.34http:0.92�1.09


historical trends in background exposures. Although uncertainty is inherent when estimating historical exposures, systemic bias was improbable. It was not possible to adjust directly for tobacco smoking, but previous adjustment in the same cohort revealed little difference in the risk for lung cancer with or without adjustment. A possible interaction between average exposure and duration of employment on mortality from lung cancer may explain some of the apparent paradoxes of the results, such as the observation that cumulative exposure, adjusted for duration of employment, had a greater effect while average exposure, adjusted for duration of employment, did not. The study provided evidence for an association between sources of EC (predominantly diesel) and the risk for lung cancer.]

2.2.4 Miners

In the cohort described in Section 2.2.1, Boffetta et al. (1988) studied 2034 men who reported working as a miner on the basis of any past employment in this occupation. The age- and tobacco smoking-adjusted relative risk for lung cancer was 2.67 (95% CI, 1.63–4.37; 15 deaths) compared with other men who did not report working as a miner and who reported no exposure to diesel exhaust. [The Working Group noted that this study was limited by the small number of miners and the lack of information regarding specific exposures to exhaust. Only 14% of miners reported exposure to diesel exhaust; these miners may also have been exposed to other lung carcinogens, such as silica or radon.]

In the cohort described in Section 2.2.1, Guo et al. (2004a) reported that male miners in three different occupational categories had an elevated risk of lung cancer. These included mine and quarry work involving metal ore (SIR, 3.26; 95% CI, 2.28–4.51; 36 cases), mine and quarry work involving non-metal ore (SIR, 1.85; 95% CI, 1.59–2.14; 181 cases) and other unspecified mine and quarry work (SIR, 1.73; 95% CI, 1.35–2.19;

70 cases). All three groups were classified by an expert review as having been exposed to diesel exhaust but not gasoline exhaust. In a separate report, Guo et al. (2004b) presented data for other cancers in male miners. In non-metal ore miners and quarry workers, the risks were elevated for leukaemia (SIR, 2.31; 95% CI, 1.39–3.61; 19 cases), oesophageal cancer (SIR, 1.74; 95% CI, 0.70–3.58; seven cases), kidney cancer (SIR, 0.88; 95% CI, 0.47–1.50; 13 cases) and urinary bladder cancer (SIR, 1.16; 95% CI, 0.73–1.76; 22 cases). Too few cases occurred in other mining groups to carry out a meaningful assessment. [The Working Group noted that these studies were limited due to the lack of detailed work histories and information on tobacco smoking. In particular, for lung cancer, these miners may have had confounding exposures to other substances, such as silica and radon.]

Neumeyer-Gromen et al. (2009) updated the mortality from lung cancer in a cohort of 5862 German underground potash miners, first described by Säverin et al. (1999), from 1970 up to 2001. Diesel equipment was introduced into potash mines in 1969 and, in 1991, the mines were closed. Tobacco smoking histories were available from medical and personnel records for 80% of the cohort. Estimates of diesel exposure were obtained in 1992, and expressed as total carbon in respirable dust. Because technology had not changed, these levels were assumed to be representative of previous exposure and were used for its categorization. The overall standardized mortality ratio was not elevated for lung cancer (SMR, 0.73; 95% CI, 0.57–0.93; 61 deaths) or urinary bladder cancer (SMR, 0.80; 95% CI, 0.40–1.60; 8 deaths). Using Cox regression modelling, the smoking-adjusted relative risk in the highest category of exposure dichotomized at 4.90 mg/m3–years was 1.28 (95% CI, 0.61–2.71; 61 cases). In a subgroup of 3335 workers who had worked underground for at least 10 years, the age-and smoking-adjusted relative risk was 1.50 (95% CI, 0.66–3.43; 37 cases). Adjusting for smoking

181

http:0.66�3.43http:0.61�2.71http:0.40�1.60http:0.57�0.93http:0.73�1.76http:0.47�1.50http:0.70�3.58http:1.39�3.61http:1.35�2.19http:1.59�2.14http:2.28�4.51http:1.63�4.37


resulted in higher risk estimates. In a model that further adjusted for time since hire and calendar year, the relative risk in the entire cohort was 2.53 (95% CI, 1.13–5.69) and that in the subcohort of workers who had worked for more than 10 years after 1969 was 3.30 (95% CI, 1.30–8.37). Using time since first hire as the time variable in a Cox regression analysis to account for duration of employment, and adjusting for age and smoking, a non-significant trend (P = 0.19) in risk for mortality from lung cancer was observed with greater exposure in the entire cohort (RR, 1.81; 95% CI, 0.92–3.58; and 1.59; 95% CI, 0.75–3.40; for the second and third tertiles, respectively). A non-significant increased trend (P = 0.17) in risk of was also found within the subcohort, for which more accurate information on exposure was available. [The Working Group noted that, although the power of the study was limited by the sample size, one of its strengths was that the effects of smoking were considered together with quantitative estimates of exposure based on measurements. The study also used an internal comparison group. Another strength in the design was the control of confounding for other mining-related occupational risk factors for lung cancer, because exposure to radon, silica dust, asbestos and heavy metals were not significant in potash mining. This study was supportive of an effect of exposure to diesel exhaust on the risk of lung cancer. ]

Attfield et al. (2012) studied the mortality of a cohort of 12 315 blue-collar workers who were employed in one of eight non-metal mines in the USA for at least 1 year after diesel equipment had been introduced. The mines were selected to minimize exposures to silica, radon and asbestos. Detailed work histories were abstracted from company records and mortality was assessed up to 1997. The dates of the introduction of diesel equipment ranged from 1947 to 1967. Historical estimates of exposure to respirable EC were constructed on the basis of personal measurements taken in the mines in 1998–2001,

which were extrapolated retroactively based on a model using diesel exhaust-related determinants, including diesel engine horse power and ventilation rates, and historical measurements of carbon monoxide. The modelled trends in concentrations of CO for previous years were then used to adjust the 1998–2001 levels of exposure to respirable EC to estimate historical annual concentrations of respirable EC for each job. Estimates of exposure to silica, asbestos, respirable dust, radon and other PAHs were also made. Estimates of exposure to diesel exhaust in surface workers were based on measurements made in 1998–2001 and no reconstruction of historical exposure was carried out. The mean exposure of surface workers only was 1.7 µg/m3 and that of the ever-underground miners was 128.2 µg/m3. Standardized mortality ratios using external referents were determined from state-specific mortality rates, and their calculation was limited to persons employed since 1960 (12 270 subjects), because state-specific rates were not available for earlier years. The standardized mortality ratio for lung cancer was 1.26 (95% CI, 1.09–1.44) for the complete cohort, 1.21 (95% CI, 1.01–1.45) for workers involved in any underground work and 1.33 (95% CI, 1.06–1.66) for workers involved exclusively in surface work. In the complete cohort, standardized mortality ratios were 1.09 (95% CI, 0.58–1.86) for urinary bladder cancer, 1.18 (95% CI, 0.76–1.74) for leukaemia, 0.98 (95% CI, 0.54–1.64) for kidney cancer, 1.12 (95% CI, 0.76–1.60) for pancreatic cancer and 0.85 (95% CI, 0.60–1.16) for prostate cancer. Mortality from oesophageal cancer was significantly elevated in all workers (SMR, 1.83; 95% CI, 1.16–2.75). In an internal analysis of the entire cohort, adjustment for the location of work (ever surface or underground) and a 15-year lag for cumulative exposure resulted in relative risks by quartile of 1.0, 0.55 (95% CI, 0.35–0.85), 1.03 (95% CI, 0.60–1.77) and 1.39 (95% CI, 0.78–2.48). In an analysis of surface workers only, the corresponding relative risks by quartiles of cumulative exposure

182

http:0.78�2.48http:0.60�1.77http:0.35�0.85http:1.16�2.75http:0.60�1.16http:0.76�1.60http:0.54�1.64http:0.76�1.74http:0.58�1.86http:1.06�1.66http:1.01�1.45http:1.09�1.44http:0.75�3.40http:0.92�3.58http:1.30�8.37http:1.13�5.69


were 1.0, 1.28 (95% CI, 0.64–2.58), 0.73 (95% CI, 0.35–1.53) and 1.00 (95% CI, 0.44–2.28). The relative risks for average exposure among surface workers were 1.0, 1.71 (95% CI, 0.82–3.58), 2.22 (95% CI, 1.01–4.90) and 2.56 (95% CI, 1.09–6.03). In an analysis of ever underground miners only, the relative risks by quartile of cumulative exposure were 1.0, 1.50 (95% CI, 0.86–2.62), 2.17 (95% CI, 1.21–3.88) and 2.21 (95% CI, 1.19–4.09), and those by quartile of average exposure were 1.0, 1.73 (95% CI, 0.99–3.05), 2.11 (95% CI, 1.14–3.90) and 1.86 (95% CI, 0.98–3.52). Further analysis of underground miners with ≥ 5 years of exposure showed similar patterns (test for trend using log cumulative exposure as a continuous variable, P = 0.015). [The Working Group noted that the log cumulative exposure best fit the exposure– response curve over the entire range. In addition, all the internal analyses of surface workers involved groups with very small ranges of exposure. Moreover, no historical exposure assessments were available for surface workers, which further limited the interpretation of the analyses of surface workers alone.]

In a nested case–control study of the above cohort of miners, Silverman et al. (2012) obtained histories of tobacco smoking, occupation and previous respiratory disease by interview for 198 lung cancer cases and 562 control subjects. Controls were assigned by random sampling from all members of the cohort who were alive before the day the case subject died and were matched on birth year, gender, ethnicity and mine. Analyses of trend in risk by level of respirable EC were adjusted for smoking, previous respiratory disease and a history of jobs that entailed a high risk of lung cancer. Smoking adjustments included separate terms for levels of smoking of surface and underground miners, to take into account that the differences in the odds ratios (ORs) for levels of smoking between these two groups. In analyses of surface and underground miners combined, the odds ratios by quartile of cumulative exposure (15-year lag)

were 1.0, 0.74 (95% CI, 0.40–1.38), 1.54 (95% CI, 0.74–3.20) and 2.83 (95% CI, 1.28–6.26; P for trend = 0.001). Corresponding odds ratios for average exposure (15-year lag) were 1.0, 1.11 (95% CI, 0.59–2.07), 1.90 (95% CI, 0.90–3.99) and 2.28 (95% CI, 1.07–4.87; P for trend = 0.062). Analyses restricted to surface miners gave odds ratios by quartile of cumulative exposure (15-year lag) of 1.0, 3.98 (95% CI, 0.69–23.02), 0.76 (95% CI, 0.12–4.98) and 0.42 (95% CI, 0.05–3.59; P for trend = 0.12), and corresponding odds ratios for average exposure (15-year lag) of 1.0, 4.38 (95% CI, 0.56–34.24), 5.67 (95% CI, 0.77–42.06) and 1.31(95% CI, 0.14–12.01; P for trend = 0.66). For underground miners, odds ratios by quartile of cumulative exposure (15-year lag) were 2.46 (95% CI, 1.01–6.01), 2.41 (95% CI, 1.00–5.82) and 5.10 (95% CI, 1.88–13.87; P for trend = 0.004), and the corresponding odds ratios for average exposure (15-year lag) were 1.0, 1.04 (95% CI, 0.45–2.43), 2.19 (95% CI, 0.87–5.53) and 5.43 (95% CI, 1.92–15.31; P for trend = 0.001). Analysis of the interaction between cumulative exposure to respirable EC and smoking showed an increased risk with increased cumulative exposure (15-year lag) for both never smokers and ever smokers (P for interaction = 0.09). Among never smokers, the odds ratio increased with increasing cumulative exposure to respirable EC from 1.47 (95% CI, 0.29–7.50; four exposed cases) for tertile 2 up to 7.30 (95% CI, 1.46–36.57; seven exposed cases). The trend of increasing risk with increased cumulative diesel exposure was attenuated in heavy smokers. [The Working Group noted that the exposure assessment methodology applied in the National Institute of Occupational Safety and Health/National Cancer Institute (NIOSH/NCI) studies was of high quality and used an established approach for occupational cohort studies. Because surface workers formed a group that had very low exposures, no significant trends were observed with increasing exposure. Consequently, the Working Group focused primarily on the combined analyses of surface

183

http:1.46�36.57http:0.29�7.50http:1.92�15.31http:0.87�5.53http:0.45�2.43http:1.88�13.87http:1.00�5.82http:1.01�6.01http:0.14�12.01http:0.77�42.06http:0.56�34.24http:0.05�3.59http:0.12�4.98http:0.69�23.02http:1.07�4.87http:0.90�3.99http:0.59�2.07http:1.28�6.26http:0.74�3.20http:0.40�1.38http:0.98�3.52http:1.14�3.90http:0.99�3.05http:1.19�4.09http:1.21�3.88http:0.86�2.62http:1.09�6.03http:1.01�4.90http:0.82�3.58http:0.44�2.28http:0.35�1.53http:0.64�2.58


and underground miners, in which the surface workers formed part of the low-exposure group. The Working Group gave the results of the case– control greater weight than those of the cohort because the former included an adjustment for tobacco smoking. The research team appeared to have used the available data most effectively. However, in any such study, uncertainties exist that may result in measurement error, and, although historical exposures might have been over- or underestimated, this could have affected the risk per unit exposure but not the pattern of the exposure–response relationship.]

2.2.5 Other groups exposed to vehicle exhausts

(a) Heavy equipment operators

Wong et al. (1985) conducted a cohort mortality study of 34 156 men who had been members of a heavy construction equipment operators union for at least 1 year between 1964 and 1978. The mortality experience of the cohort was compared with that of white men in the USA. Historical environmental measurements were not available, and only partial work histories were accessible for some cohort members. Mortality from respiratory cancer for the whole cohort was similar to that expected (SMR, 0.99; 95% CI, 0.88–1.10), with no trend by duration of union membership. A significant excess of mortality from lung cancer was found among the 4075 retirees, after excluding early retirees who were thought potentially to have retired due to illness (SMR, 1.30; 95% CI, 1.04–1.61). The standardized mortality ratio for urinary bladder cancer was 1.18 (95% CI, 0.78–1.72), with no trends by duration or latency. Data on cigarette smoking were available for a small sample of 107 workers, and did not indicate any difference in smoking habits between the cohort and the general population. [The Working Group noted that the main limitation of this study with regard to diesel exhaust was the lack of any information

documenting such an exposure; in general, it is not known to what degree heavy equipment operators have appreciable exposure to diesel exhaust.]

In the study described in Section 2.2.1, Boffetta et al. (1988) studied 855 men who reported working as heavy equipment operators. The age- and tobacco smoking-adjusted relative risk for lung cancer was 2.60 (95% CI, 1.12–6.06) compared with men who reported no such employment and no exposure to diesel exhaust. [The Working Group noted that this study was limited by the small number of heavy equipment operators included and the categorization of exposure from self-reporting. Less than half (46%) of the heavy equipment operators reported exposure.]

In the study described in Section 2.2.3, Järvholm & Silverman (2003) found that 14 364 heavy

2. cancer in humans...2. cancer in humans 2.1 introduction 2.1.1 general aspects diesel and gasoline...

Documents